Severe, exacerbation-prone asthma impacts 5-10% of the asthma population and continues to have substantial human and economic impact, with nearly 2 million emergency room visits and 0.5 million hospitalizations per year. Evidence from the Severe Asthma Research Program (SARP) supports the heterogeneity of severe asthma, with substantial evidence to suggest differentiation of these patients into 2 broad categories based on biomarker evidence for the presence/absence of Type(T)-2 (IL-4, 5, -13) associated inflammation. Concurrent industry sponsored clinical trials have further supported this broad differentiation, with evidence for substantial efficacy of T2-targeted biologic therapies including those targeted to IL-4/13 and IL-5 pathways in T2Hi asthma patients. However, the best biomarkers to predict response to T2-targeted therapies are not yet clear. Given their enormous costs, it is critical to better understand and identify those who most need these medications, which ones to utilize first and in which patients. It is even more unclear whether specific biomarkers in patients with no (using current biomarker) evidence for T2 inflammation exist or whether they predict targeted biologic approaches for these patients. The adaptive design trial proposed here will utilize currently accepted biomarkers, as well as additional exploratory bio-imaging and genetic markers to predict the most efficacious and safe approaches for these broad (but then more specific) T2-phenotypes. We therefore hypothesize that an adaptive trial design integrating T2 (and non-T2) biomarkers, targeted therapies and clinically relevant outcomes will improve the understanding of the pathobiology of severe asthma patients on medium to high dose inhaled corticosteroids (ICS), with or without long acting ??2 agonists (LABA) phenotypes and bring the most efficacious (and safest) medication to each severe asthma patient. We propose a multiphase adaptive trial design in 800 poorly controlled, exacerbating and/or severe asthmatic and/or oral corticosteroids (OCS). The 1st (run-in) phase will establish each participant's baseline over a 3-6 month period of time, while repeatedly measuring established and exploratory biomarkers. The data from this run-in phase will be used to assign the patient to a T2Hi or -Lo molecular phenotype and inform the modeling of predictive factors to be applied during the targeted treatment phase. The targeted treatment phase will consist of 3 treatments, adaptively applied to the two broad T2 phenotypes, with the intention to support the importance of potential T2 sub-phenotypes, such as a T2Hi/Mast cell-Hi and T2Lo/Metabolic. The 3 treatments will differ by starting T2 phenotype, but the primary endpoint for each intervention will be treatment failure defined by a biomarker and clinical index. T2Hi interventions will sequentially include a CRTH2 antagonist, an anti-IL-4Receptor(R) antibody and a soluble TNF-? receptor, while T2Lo interventions will include anti-IL-4R, an anti-IL-6/6Rreceptor and bronchial thermoplasty. These studies will greatly expand on the precision medicine pathway to improve the care of severe, exacerbation-prone asthma.